The 2026 MVP Races, Updated: Two Runaways and the Case for Doubting Them
Updated August 22, 2026. This is a rerun of the July 9 break edition, which is kept unchanged so the calls graded below can be checked against what it actually said. Companion to the Cy Young piece. Stat lines are through games of August 21; projections and MVP odds are from the August 22 simulator run over the 481 games still to be played.
BBMI Research DeskUpdated August 22, 2026Model run: August 22 · 500 simulated seasons
Six weeks ago the model read the American League as a two-man fight and the National League as a coronation. It now reads both leagues as coronations. Yordan Alvarez wins the AL value crown in 500 of 500 simulated finishes, Shohei Ohtani the NL in 498 of 500. That is a bigger claim than it sounds, and a smaller one, and this piece is mostly about the difference.
Bigger, because it is not a hedge. With roughly 32 games left per club, about four-fifths of the season is banked, and a 32-run lead in offensive value does not evaporate in a month. Smaller, because of what the number is not. It is not a prediction of the vote. It does not model the leader getting hurt, a caveat with unusual teeth this year, as you will see. And it is graded on offense plus two-way pitching, with no credit for defense or baserunning. That last omission is why the American League's best defender appears nowhere on the board below, and why the National League's best defender is priced at two-tenths of one percent. Read the odds as "who has produced and will produce the most value at the plate," not as "who will be handed the trophy."
Two leaders, two very different margins
Projected full-season runs above average. Bar length is value; the number at right is the share of 500 simulated finishes in which he leads BBMI's MVP value metric.
American League
Alvarez
+71.9
100.0%
Caminero
+39.5
0.0%
Rice
+29.2
0.0%
W. Contreras
+29.1
0.0%
Alonso
+28.5
0.0%
Murakami
+28.3
0.0%
National League
Ohtani
+70.8
99.6%
Wood
+45.8
0.2%
Crow-Armstrong
+39.1
0.2%
Soto
+37.4
0.0%
Harper
+33.0
0.0%
Olson
+32.2
0.0%
Bars share one scale across both leagues. Ohtani's bar is split: blue is his bat, worth +48.6 runs, and green is the 85⅔ innings he has already thrown, worth +22.2 runs saved. Value is measured against the league's average hitter per plate appearance. No defense or baserunning is included. BBMI season simulator, August 22 run.
American League: the two-man race lost a man, then lost another
In July this was Alvarez at 56.5% and Nick Kurtz at 35.5%, with Aaron Judge lurking as the most underrated bat in the sport. Alvarez is now at 100%, and the honest accounting is that he earned roughly half of that and was handed the other half. He has been the best hitter in the league since the break. He has also watched both of his challengers land on the 60-day injured list.
In July we wrote that his was "the rare hot start that projects to get better," because a .708 expected slugging sat above a .621 actual with no regression siren anywhere in the batted-ball data. Six weeks on, the line has gone up: .321/.434/.618, a 1.052 OPS, 36 homers, 90 RBI. The gap is still open, too, with a .673 xSLG against .618 and a .458 expected wOBA against a .435 actual. He is one of only three hitters in baseball the model projects past 100 runs and 100 RBI, at 43 HR, 108 RBI and 106 R (Ohtani and Pete Alonso are the others), and his +71.9 projected runs above average is 82% clear of the next AL bat. The team case has improved as well: Houston is 65-64 but leads the AL West race in the model's eyes, at 52% to win the division and 60% to reach October. The only things that beat him now are the calendar and the training room.
Junior Caminero3B · Rays
Best of the field · 0.0%
ActualSlash .276/.363/.536HR 35xSLG .498OAA −17
ProjectedSlash .272/.357/.527HR 44RBI 104R 98SB 4
The July piece flagged mild regression here, on expected stats a touch under the actuals, and that is exactly what arrived: .288/.378/.561 then, .276/.363/.536 now, with the expected slugging still trailing by 38 points. He remains the AL's second-most valuable bat by a comfortable margin and the model's projected AL home run leader at 44, on the league's best team. Tampa Bay is 76-52, a lock for October and about 30% to reach the World Series. What the odds column can't show is that a 44-homer season on a division winner is a real down-ballot MVP case even at 0%. The model isn't saying he'll finish tenth; it's saying he will not out-produce Alvarez. One number to sit with: −17 outs above average at third base, the worst mark among everyone in this article. Our metric ignores defense. The voters won't.
Projectednone. The model projects no further 2026 games; the line at left is carried forward unchanged.
The July call was blunt: "the field's biggest regression risk among the leaders," on a .388 BABIP, a 31.3% strikeout rate, and expected stats sitting under his actuals. The regression showed up on schedule. He was at .275/.415/.512 with 20 homers on July 9 and .256/.396/.483 with 21 on July 31, and then the injury ended the argument. He has not played since, and he is on the 60-day injured list. His frozen line is still worth +25.4 projected runs above average, seventh in the league, which tells you how good the first four months were. It also means the model's July 35.5% was, in hindsight, pricing a player who had about three weeks of season left in him.
Aaron JudgeRF · Yankees
60-day IL · last played May 31
ActualSlash .248/.375/.533HR 17Games 59PA 261
Projectednone. The model projects no further 2026 games; the line at left is carried forward unchanged.
This is the one to mark plainly. In July we called Judge "the field's top positive-regression bet for the second half," pointing at a .286 BABIP, a .600 xSLG against a .533 actual, the best barrel rate in the group, and 6% MVP odds we argued were a playing-time discount rather than a talent one. The playing-time discount was the whole story. He has not appeared in a game since May 31, he is on the 60-day injured list, and he no longer carries enough plate appearances to qualify for a Statcast leaderboard, let alone a ballot. The model's low number was right. Our framing of it, that the underlying data made him a buy, was a bet on a return that never came. A rate-luck signal is worthless without the health to spend it.
The rest of the AL board is mostly names the July piece did not have room for. Ben Rice (Yankees, 34 HR, .530 slugging) turned the power breakout we flagged into a full season and is now New York's best bat by default. Willson Contreras, hitting .285/.396/.541 for a 69-59 Boston club that is 98% to play in October, and Pete Alonso, with 30 homers, 85 RBI and a .544 xSLG that says he has been unlucky in Baltimore, sit within a run of each other. And Munetaka Murakami is the most interesting name in the league: a .225 average that hides a 20.6% barrel rate and a 58.8% hard-hit rate, both of which lead the American League among hitters with a meaningful batted-ball sample, with 29 homers in 93 games for a Chicago White Sox team that is, improbably, 67-61 and 80% to win the Central.
The luck board, six weeks later
Actual slugging (dark dot) against Statcast expected slugging (colored dot)
.440.500.560.620.690
Out-hitting his results, so expect a rebound
Alvarez
.618 to .673
Wood
.535 to .585
Soto
.539 to .589
Alonso
.504 to .544
Witt Jr.
.463 to .509
Out-running his contact, so expect a give-back
Crow-Armstrong
.554 to .470
Abrams
.530 to .446
Caminero
.536 to .498
Schwarber
.522 to .491
Red dots AL, blue dots NL. Crow-Armstrong and Caminero were both flagged as negative-regression risks in the July edition and both still sit above their expected marks. Crow-Armstrong and Abrams share the widest gaps on this board at 84 points of slugging each. Statcast batted-ball data through August 21.
National League: Ohtani leads on a season that already ended
The July piece put Ohtani at 95% and gave one reason: MVP rewards total value, and he was the only man adding a second position. That is still the reason. It is also, now, entirely in the past tense, and that is the single most interesting thing on this board.
Take the two halves separately, because the model does. The bat: .291/.390/.552 with 30 homers, projecting to 38 HR, 101 RBI and 108 R, and a .409 expected wOBA that sits above his .395 actual, so there is no mirage to correct. That is worth +48.6 runs above average, the most in the league. The arm: 14 starts, 85⅔ innings, an 8-2 record and a 1.79 ERA, worth another +22.2 runs saved against league-average run prevention. Together, +70.8, a 25-run cushion over the field, and the reason this is 99.6% rather than a coin flip with James Wood.
Here is the part that changes the story. Ohtani has not pitched since July 3. He is on the active roster and was in the lineup as recently as August 21, but he has not taken the mound in seven weeks, and the model projects him zero further innings because he is not in the Dodgers' current rotation. So the two-way case is now a bank balance rather than an income stream. Strip the arm out and his bat alone, at +48.6, leads Wood's +45.8 by under three runs, which is a genuine race. The 99.6% is the model saying that 85⅔ innings of 1.79 baseball, already earned, is more than the field can make up in the season's last month. It is not saying he is still doing it.
In July we called him "the loudest positive-regression signal in the field" and projected a 40-homer, 130-run season. He delivered on the shape of it, with 30 homers and 100 runs in 114 games, and he is still the model's projected runs-scored leader at 123. The contact quality is, if anything, louder: a .585 xSLG against a .535 actual, a 20.8% barrel rate, a 57.6% hard-hit rate. He is the second-most valuable bat in the National League and the best one that isn't also a pitcher. Two things cap him. He has been on the injured list since August 3, and the model assumes short-IL players return and gives him the rest of the schedule, which is an assumption worth naming. And Washington ended the July piece seven back at a fringe 17% and has since fallen to 61-69 and 0%. The ceiling we wrote about is real. The stage isn't there.
The most instructive name in either league, because he is the clearest case of our metric failing a player. In July we flagged him as "the field's clearest negative-regression risk," and on contact quality that call has held and then some: a .470 expected slugging against a .554 actual is the widest such gap on our board, tied with his fellow National Leaguer CJ Abrams, and a .341 BABIP is still doing work. But he has gone from 19 homers and 23 steals to 32 and 31, on a 74-55 Cubs team that is 98% to play in October, while posting +23 outs above average in center field. Our MVP number is offense only. Add elite center-field defense and 31 bases and his real case is materially better than the +39.1 runs and 0.2% we publish. If a human ballot goes anywhere other than Ohtani, this is the most likely place.
The July verdict was "the second-best bat, wasted." Everything has gotten worse except the bat. He still walks far more than he strikes out, 17.5% against 13.4%, and his .589 expected slugging still says he has been underpaid for his contact. But he has not played since July 24, and the Mets have gone from 40-54 to 58-71, last in the East, 0% to reach October, on pace for a season nobody in Queens will want to revisit. In 84 games he is still the fourth-most valuable bat in the league. That is the whole tragedy in one number.
Also on the board:Bryce Harper, at .257/.379/.488 with a 16.7% walk rate and +33.0 runs, has quietly become Philadelphia's most valuable hitter, ahead of teammate Kyle Schwarber. Schwarber leads the majors with 37 home runs and projects to 47, but a .242 average, a 33.6% strikeout rate and a .491 xSLG keep his all-around value below his highlight reel. Matt Olson, with 36 homers and 44 projected, is Atlanta's engine on a 75-54 club. And Freddie Freeman is hitting .300 again for a 78-51 Dodgers team that is the World Series favorite. He is reliably excellent, and, as in July, a complement to Ohtani rather than a rival.
Where they stand, and where the model lands them
Every hitter named above: banked 2026 line through August 21, against the median full-season projection across 500 sims
Actual, through Aug 21
Projected full season
Player
G
AVG/OBP/SLG
HR
RBI
R
SB
AVG/OBP/SLG
HR
RBI
R
SB
Yordan AlvarezHOU
127
.321/.434/.618
36
90
85
1
.311/.421/.591
43
108
106
1
Junior CamineroTB
128
.276/.363/.536
35
81
79
3
.272/.357/.527
44
104
98
4
Ben RiceNYY
123
.249/.345/.530
34
80
78
3
.247/.339/.521
42
99
98
4
Willson ContrerasBOS
117
.285/.396/.541
26
78
65
3
.280/.368/.524
32
97
83
4
Pete AlonsoBAL
129
.271/.361/.504
30
85
79
5
.269/.354/.498
37
104
100
6
Munetaka MurakamiCWS
93
.225/.362/.526
29
58
68
1
.230/.359/.509
37
79
91
2
Nick Kurtz 60-day ILATH
99
.256/.396/.483
21
69
62
7
no further games projected
Bobby Witt Jr.KC
111
.288/.363/.463
16
48
71
33
.286/.360/.461
20
63
90
40
Aaron Judge 60-day ILNYY
59
.248/.375/.533
17
38
43
5
no further games projected
Shohei OhtaniLAD
123
.291/.390/.552
30
80
84
8
.284/.385/.544
38
101
108
10
James WoodWSH
114
.265/.393/.535
30
73
100
17
.262/.388/.524
37
92
123
20
Pete Crow-ArmstrongCHC
129
.279/.378/.554
32
80
91
31
.272/.360/.537
39
99
113
37
Juan SotoNYM
84
.283/.408/.539
21
52
46
7
.274/.404/.519
28
73
65
10
Bryce HarperPHI
129
.257/.379/.488
25
72
81
7
.253/.377/.478
31
90
99
9
Matt OlsonATL
129
.261/.337/.539
36
77
89
4
.259/.337/.531
44
98
108
4
Kyle SchwarberPHI
123
.242/.361/.522
37
74
73
3
.238/.353/.515
47
94
96
4
Freddie FreemanLAD
125
.300/.372/.463
15
61
62
6
.297/.372/.460
19
79
80
7
CJ AbramsWSH
124
.279/.354/.530
29
91
77
24
.275/.343/.518
35
111
95
28
Team colored red for AL, blue for NL. Projections are the median of 500 replays of the remaining schedule, added to what is already banked, so a projected line can only move up from the actual one. Kurtz and Judge are on the 60-day injured list and are projected no further games, which is why their season totals are already final. BBMI season simulator, August 22 run.
What our MVP number leaves out
Outs above average, 2026. Defense is not in the model's MVP value at all.
Witt Jr. · SS
+24
Crow-Armstrong · CF
+23
Olson · 1B
+7
Ohtani · DH
0
Alvarez · DH
−2
Wood · RF
−6
Abrams · SS
−17
Caminero · 3B
−17
Red AL, blue NL; the tick is league average. Our MVP value is bat-only, plus pitching for two-way players, so a 41-out swing between Witt and Caminero is invisible to the odds above. Bobby Witt Jr. does not appear on the AL value board at all. Both of our leaders are designated hitters, which means neither is being credited, or charged, for anything with a glove. Statcast outs above average through August 21.
Grading the July board
We publish these boards so they can be checked later, so here is the check. Five calls from the July 9 edition, and what six weeks did to them.
Alvarez: "no regression siren here." He was out-hitting a .621 slugging with a .708 expectation. He now slugs .618 with a .673 expectation, and his odds went from 56.5% to 100%.
Held
Kurtz: "the field's biggest regression risk." A .388 BABIP and a 31.3% strikeout rate said the on-base was partly air. His line fell from .275/.415/.512 to .256/.396/.483 in three weeks, and then he was hurt.
Held
Judge: "the field's top positive-regression bet." The batted-ball read was sound and completely beside the point. He has not played since May 31 and is on the 60-day injured list. A luck signal you can't spend isn't a signal.
Missed
Crow-Armstrong: "the clearest negative-regression risk." The gap between his results and his contact is wider now, not narrower, but he answered with 13 more homers and eight more steals, and his team is 98% to make the playoffs. Right on the mechanism, wrong on the consequence.
Split
Wood: "the loudest positive-regression signal," projecting 40 HR and 130 runs. He hit 30 homers and scored 100 in 114 games with an even louder contact profile, and still projects to 123 runs. The bat came through. The injury and the standings did not.
Held
The contender test
Playoff probability of each candidate's team, August 22 run
Dodgers · Ohtani, Freeman
100%
Rays · Caminero
100%
Yankees · Rice
100%
Braves · Olson
99%
Cubs · Crow-Armstrong
98%
Red Sox · W. Contreras
98%
White Sox · Murakami
90%
Phillies · Harper, Schwarber
83%
Astros · Alvarez
60%
Orioles · Alonso
7%
Nationals · Wood, Abrams
0%
Mets · Soto
0%
Red bars AL, blue bars NL. Alvarez's 60% is the softest team case among the leaders, but Houston also leads the AL West in the model at 52% to win the division, so his October story is still being written. Athletics (Kurtz) omitted at 0%. BBMI playoff simulator, 10,000 runs, August 22.
The bottom line
Both BBMI value races are effectively decided in the model, for different reasons. Alvarez leads the American League board by being the best hitter in it and then outlasting a field that broke; a 1.052 OPS is the honest half of a 100% that injuries wrote the rest of. Ohtani finishes first in BBMI's NL value metric on a two-way résumé he stopped adding to on July 3, which makes his 99.6% a statement about arithmetic already banked rather than a player still compounding.
The interesting reading is in the gaps. Strip the pitching out and Ohtani's lead over Wood is under three runs. Add defense and baserunning back in, which our number does not do, and Pete Crow-Armstrong's case moves more than anyone's. And if you want one bat to watch in September, it is Munetaka Murakami, hitting .225 with the loudest contact profile in the American League for a White Sox team nobody expected to be leading a division in late August.
We will run this board again at the end of the regular season, and grade this one the same way.
Methodology: slash lines and counting stats from official box-score data through August 21. MVP probabilities and full-season projections are BBMI season-simulator outputs from the August 22 run: each player's projected line is his banked 2026 total plus a Monte Carlo distribution over the 481 remaining games, replayed 500 times, with the median reported. MVP value is runs above average from a linear-weights batting model, meaning rate times playing time, measured against the league's average hitter per plate appearance, plus, for two-way players, runs saved on the mound versus league-average run prevention. Defense and baserunning are not included. The odds are the share of the 500 simulated finishes in which a player leads his league in that value. They are not a forecast of the writers' vote, and they do not model the leader being injured. Players who are done for the season keep their banked line and are projected no further games; players on the short-term injured list are assumed to return. Quality-of-contact figures (xSLG, xwOBA, barrel rate, hard-hit rate) and outs above average are Statcast batted-ball data through August 21. Team records and playoff probabilities are from the August 22 BBMI playoff simulator, 10,000 runs. For the pitching companion, see the 2026 Cy Young races. Not betting advice.